https://github.com/google-research/look-back-when-surprised
Look Back When Surprised: Stabilizing Reverse Experience Replay for Neural Approximation
https://github.com/google-research/look-back-when-surprised
reinforcement-learning replay-buffer
Last synced: about 1 year ago
JSON representation
Look Back When Surprised: Stabilizing Reverse Experience Replay for Neural Approximation
- Host: GitHub
- URL: https://github.com/google-research/look-back-when-surprised
- Owner: google-research
- License: mit
- Created: 2022-07-01T09:44:10.000Z (almost 4 years ago)
- Default Branch: main
- Last Pushed: 2025-03-22T01:53:37.000Z (about 1 year ago)
- Last Synced: 2025-04-03T01:02:37.752Z (about 1 year ago)
- Topics: reinforcement-learning, replay-buffer
- Language: Python
- Homepage:
- Size: 105 KB
- Stars: 10
- Watchers: 2
- Forks: 2
- Open Issues: 26
-
Metadata Files:
- Readme: README.md
- Contributing: CONTRIBUTING.md
- License: LICENSE
Awesome Lists containing this project
README
# Look Back When Surprised: RER++
This repository serves to open-source the code used in the paper: "[Look Back When Surprised: Stabilizing Reverse Experience Replay for Neural Approximation](https://arxiv.org/abs/2206.03171)".
### Getting started
To avoid any conflict with your existing Python setup, it is suggested to work in a virtual environment with [`virtualenv`](https://docs.python-guide.org/dev/virtualenvs/). To install `virtualenv`:
```bash
pip install --upgrade virtualenv
```
Create a virtual environment, activate it and install the requirements in [`requirements.txt`](requirements.txt).
```bash
virtualenv env
source env/bin/activate
pip install -r requirements.txt
```
### Training & Testing
Our models can be trained as follows:
```bash
python -W ignore -m src.main --exp_name --algo --replay_buffer_sampler --env --train --seed $1 --snapshot_dir $2
```
# Paper Citation
If you find our codes useful, do consider citing our [paper](https://arxiv.org/abs/2206.03171):
```
@article{kumar2022look,
title={Look Back When Surprised: Stabilizing Reverse Experience Replay for Neural Approximation},
author={Kumar, Ramnath and Nagaraj, Dheeraj},
journal={arXiv preprint arXiv:2206.03171},
year={2022}
}
```
# References
Our repository makes use of various open-source codes. Most of which have been documented at Garage. If you find the respective codes useful, do cite their respective papers as well:
```
@misc{garage,
author = {The garage contributors},
title = {Garage: A toolkit for reproducible reinforcement learning research},
year = {2019},
publisher = {GitHub},
journal = {GitHub repository},
howpublished = {\url{https://github.com/rlworkgroup/garage}},
commit = {be070842071f736eb24f28e4b902a9f144f5c97b}
}
```
# Disclaimer
**This is not an officially supported Google product.**